Search results
A total of 165 records from electronic databases such as MIDLINE/PubMed, Science Direct, Hinari, Cochrane Library, CINAHL, and African Journals Online, Google, and Google Scholar were searched; however 25 records were excluded due to duplication. We excluded 93 records because these articles were not related to our review after assessing their titles and abstracts. After assessing 47 full articles, 27 articles were further excluded for reasons (the outcome variables were not reported). Finally, 20 studies were included in this systematic review and meta-analysis (Figure 1).
Characteristics of the included articles
The current systematic review and meta-analysis included 20 primary studies with a total of 93,600 study participants. Of the total of 20 studies; 7 studies were conducted in Amhara region [10, 31-36], 5 studies in SNNP region [26, 37-40], 2 studies in Oromia region [9, 41], 2 studies in Addis Ababa city administration [25, 42], 1 study in Tigray region [43], 1 study in Dire Dawa city administration [44], 1 study in Harar region [45] and 1 study was conducted national-wide [46] (Table 1). Only 4 studies were conducted in the rural area, the rest 16 were conducted in the urban area. Among the 20 primary studies included in this meta-analysis, 18 studies were conducted in a health institution and only two studies were community-based. Three studies of a total of 20 studies were used in a case-control study design, and the rest were used in a cross-sectional study design. The sample size of the primary studies included in this systematic review and meta-analysis ranged from 144 to 68,437 as reported by Debre tabor in the Amhara region and national-wide studies respectively [34, 46]. The highest magnitude of post-partum hemorrhage was 16.6% in SNNP region [26], and the lowest report was 1.4% in Addis Ababa [25].
Table 1
Summary of the studies included in the systematic review and meta-analysis of magnitude and associated factors of post-partum hemorrhage in Ethiopia 2020.
Authors
|
Publication year
|
Region
|
Sample size
|
Magnitude
(%)
|
Study designs
|
Area
|
Study setting
|
Sampling technique
|
Mesfin et al [43]
|
2021
|
Harar
|
642
|
12.9
|
Cross-sectional
|
Urban
|
Institution-based
|
Simple random
|
Hamdela B. Et al [35]
|
2021
|
SNNPR
|
2517
|
2.9
|
Cross-sectional
|
Urban
|
Community-based
|
Simple random
|
Worku T. Et al [29]
|
2019
|
Amhara
|
384
|
11
|
Cross-sectional
|
Urban
|
Institution-based
|
Systematic random
|
Adere A. Et al [40]
|
2020
|
Addis Ababa
|
606
|
12.4
|
Case control
|
Urban
|
Institution-based
|
Systematic random
|
Abebaw A. Et al [30]
|
2020
|
Amhara
|
424
|
14.2
|
Cross-sectional
|
Urban
|
Institution-based
|
Systematic random
|
Sufian Jeilu., et al [42]
|
2021
|
Dire Dawa
|
239
|
13
|
Cross-sectional
|
Urban
|
Institution-based
|
Systematic random
|
Legesse et al [9].
|
2017
|
Oromia
|
600
|
7
|
Case control
|
Urban
|
Institution-based
|
Simple random
|
KEBEBUSH ABERA [23]
|
2014
|
Addis Ababa
|
12995
|
1.4
|
Cross-sectional
|
Urban
|
Institution-based
|
Consecutive sampling
|
Temesgen MA [31].
|
2017
|
Amhara
|
377
|
5.8
|
Cross-sectional
|
Urban
|
Institution-based
|
Systematic random
|
Kebede BA et al [24]
|
2019
|
SNNPR
|
422
|
16.6
|
Cross-sectional
|
Urban
|
Institution-based
|
Consecutive sampling
|
TESHAGER DERBEW[38]
|
2017
|
SNNPR
|
1786
|
5.9
|
Cross-sectional
|
Rural
|
Institution-based
|
Consecutive sampling
|
Harrison MS et al [37]
|
2021
|
SNNPR
|
998
|
2.2
|
Cross-sectional
|
Urban
|
Institution-based
|
Consecutive sampling
|
Abate T. Et al [10]
|
2014
|
Amhara
|
203
|
14.8
|
Cross-sectional
|
Urban
|
Institution-based
|
Sample random
|
Geleto et al [44]
|
2020
|
National
|
68437
|
4.2
|
Cross-sectional
|
Urban
|
Institution-based
|
Survey
|
Habitamu et al [32]
|
2019
|
Amhara
|
144
|
7.6
|
Cross-sectional
|
Urban
|
Institution-based
|
Systematic random
|
Gudeta TA et al [39]
|
2018
|
Oromia
|
200
|
9.7
|
Cross-sectional
|
Urban
|
Institution-based
|
Systematic random
|
Maeruf H, et al [41]
|
2020
|
Tigray
|
752
|
3.1
|
Cross-sectional
|
Urban
|
Institution-based
|
Systematic random
|
Zimmerman et a [36]
|
2019
|
SNNPR
|
321
|
12.5
|
Cross-sectional
|
Rural
|
Community-based
|
Survey
|
Tiruneh B et al [33].
|
2020
|
Amhara
|
1060
|
9
|
Cross-sectional
|
Rural
|
Institution-based
|
Systematic random
|
Dagne AH et al [34].
|
2021
|
Amhara
|
493
|
13.6
|
Cross-sectional
|
Rural
|
Institution-based
|
Systematic random
|
SNNPR- South Nation Nationality and People Region |
Meta-analysis
A random effect model was used to estimate the pooled magnitude of postpartum hemorrhage and the factors associated with it in Ethiopia. The pooled magnitude of postpartum hemorrhage in Ethiopia was 8.18% [(95% CI; 6.996 - 9.363), I2 = 97.9%, P < 0.01] (Figure 2). Significant publication bias was observed since the Egger’s test result is significant (P = 0.014) at 95% CI. The Duval and Tweedie trim and fill methods were used to estimate the number of studies missed from a meta-analysis as a source of publication bias but the finding was not significant [47]. We also observed a symmetrical distribution of the funnel plot, indicating a significant publication bias was observed (Figure 3).
Subgroup analysis
Subgroup analysis was performed to identify the source of heterogeneity since I2 test (I2= 97.9%, P < 0.001) shows the presence of significant heterogeneity. Therefore, subgroup analysis was done using the study area region and the study setting (urban vs rural) using random model effect analysis. Accordingly, the highest pooled magnitude of postpartum hemorrhage was in the Amhara region (10.6% (95% CI: 8.14-13.18), I2 = 80.2%) and the lowest was in Addis Ababa (6.8% (95% CI: 3.96-17.60), I2 = 98.5) (Figure 4). The pooled magnitude of postpartum hemorrhage was slightly higher in rural areas 9.99% (95% CI; 6.53 -13.43), I2 = 91.2%) than in urban areas 7.69% (95% CI; 6.41-8.96), I2 = 98.1%) (Figure 5).
Factors associated with postpartum hemorrhage
This systematic review and meta-analysis identified different factors associated with postpartum hemorrhage in Ethiopia. Variables reported as statistically significant with the occurrence of postpartum hemorrhage in at least two primary studies were incorporated into this metal analysis. Accordingly, age of the women, prolonged labor, antenatal care (ANC) visits, parity, previous postpartum hemorrhage were significantly associated with postpartum hemorrhage.
The age of the mother was significantly associated with postpartum hemorrhage among two primary studies included in this systematic review and meta-analysis [26, 34]. A total of 566 study participants were included to determine the association between the age of the mothers and the occurrence of post-partum hemorrhage. The pooled odds ratio showed that older mothers (>35 years) had 5 times more odds of developing postpartum hemorrhage than mothers younger than 35 years [OR = 5.038 (95% CI; 2.774 - 9.151), I2 = 0.0%, P = 0.602] (Figure 6).
Prolonged labor was reported as a factor associated with postpartum hemorrhage among two primary studies [33, 37]. To analyze the association between prolonged labor and the postpartum hemorrhage, 2,894 mothers were included in the meta-analysis. Accordingly, the odds of postpartum hemorrhage were 4 times higher in mothers who had prolonged labor compared to mothers whose labor had not been prolonged [OR = 4.054 (95% CI; 1.484 - 11.074) I2 = 70.8%, P = 0.064] (Figure 7).
Two primary studies have reported that antenatal care follow-up (ANC) reduces the occurrence of post-partum hemorrhage after birth [33, 34]. More than 500 postnatal mothers were included to show the association between antenatal care follow-up and postpartum hemorrhage. The pooled result showed that mothers who didn’t had ANC follow-up were 13.8 times more likely to develop postpartum hemorrhage than mothers who had ANC follow-up [OR = 13.84 (95% CI; 5.57 - 34.346) I2 = 21.9%, P = 0.258] (Figure 8).
Parity was reported as a factor associated with postpartum hemorrhage among two primary articles included in this meta-analysis [33, 34]. A total of 521 mothers were included to analyze the association between parity and postpartum hemorrhage. The odds of postpartum hemorrhage among grand multipara mothers were 6.6 times higher than multi- and Primi-Para mothers [OR = 6.584 (95% CI; 1.902 - 22.795) I2 = 58.8%, P= 0.119] (Figure 9).
Two primary studies included in this meta-analysis reported that pervious postpartum hemorrhage was significantly associated with the occurrence of post-partum hemorrhage after birth [26, 34]. A total of 566 mothers were included to show the association between previous PPH and postpartum hemorrhage. According to the analysis, mothers who had postpartum hemorrhage in the previous delivery were 4 times the risk of developing postpartum hemorrhage than mothers who didn’t have postpartum hemorrhage previously [OR = 4.355 (95% CI; 2.347 - 8.079) I2 = 0.0%, P = 0.457] (Figure 10).